PROJECT SUMMARY/ABSTRACT Pancreatic cancer is a rapidly lethal cancer and accounts for the fourth highest number of cancer-related deaths in the US. Few risk factors for this fatal cancer are well-established and identifying new susceptibility markers for this disease may provide new avenues for research. In a recent study, DNA methylation markers directly measuring regulatory T-lymphocytes were strongly associated with subsequent colorectal and lung cancer risk in a prospective cohort study (Barth et al., JNCI 2015). In addition, we have shown that profiles of DNA methylation in leukocytes are associated with human bladder and head and neck cancers (both tobacco- related cancers). Importantly, we have demonstrated that much, but not all, of this variation in DNA methylation linked to these diseases is attributable to disease-associated changes in the underlying profile of normal, unactivated immune cells in the blood. Critical to this application, we have developed and validated a novel statistical methodology that can utilize our independently generated library of DNA methylation profiles from leukocyte subsets to quantify the proportion of these normal leukocyte subtypes in archival material. This method can disentangle the variability in DNA methylation profiles attributable to variation in leukocyte sub- populations from potentially important and novel epigenetic variation arising within a cell population. The well- described risk associated with inflammatory factors in pancreatic cancer provides a compelling rationale for undertaking a prospective analysis characterizing epigenetic biomarkers of the immunologic perturbations. Our approach offers a unique setting in which to discover induced immunomodulatory states in the blood that are currently not recognized as risk factors for pancreatic cancer. We hypothesize that potentially causal immunologic perturbations characterized by changes in blood cell proportion and immune activation (measured with DNA methylation markers) will be prospectively discernible in the peripheral blood. To test our hypothesis, we propose to examine DNA methylation changes in relation to risk of pancreatic cancer using existing samples from four large prospective cohort studies.